Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Evolutionary Ecology 13: 605±618, 1999. Ó 2001 Kluwer Academic Publishers. Printed in the Netherlands. Evolutionary perspective Experimental approaches to studying the initial evolution of conspicuous aposematic signalling LEENA LINDSTROÈM Department of Biological and Environmental Science, University of JyvaÈskylaÈ, PO Box 35, FIN-40351 JyvaÈskylaÈ, Finland (tel.: +358 14 260 4197; fax: +358 14 260 2321; e-mail: [email protected].®) Received 9 June 2000; accepted 27 November 2000 Co-ordinating editor: C. Rowe Abstract. Aposematism, where prey species conspicuously advertise their unpro®tability to predators, is a widespread defensive strategy. One feature of an aposematic anti-predatory strategy that is especially puzzling is conspicuousness. While conspicuousness aids associative learning in predators, it involves being more visible, which probably increases predation risk. Although aposematism is an old evolutionary question, experimental studies to its evolution have been scarce. Only 11 experiments address the potential bene®ts of conspicuousness, which have successfully manipulated conspicuousness. This is probably because it is dicult to separate conspicuousness from other characters of aposematic prey, e.g. colour. Furthermore since predators and prey species have coexisted for a long time, and there might be special adaptations other than conspicuous signalling, our experimental results might be confounded with, e.g. predatory biases. In this review, I will examine the problems of studying the costs and bene®ts of conspicuousness as well as the initial evolution of conspicuousness and the recent progress in the study of aposematism. Key words: aposematism, initial evolution, signal conspicuousness Evolutionary paradox The theory of aposematism, where prey species advertise their unpro®tability to predators by conspicuous means (for example, through the use of colours, odours and behaviour), dates back to Wallace (1867), and has remained an attractive evolutionary topic ever since (Rowe and Guilford, 2000). The idea that Wallace introduced was that being distasteful was not sucient as an antipredatory defence for butter¯y larvae, since sampling by predators might fatally injure them. However, if caterpillars were conspicuously coloured and looked as distinct as possible from palatable prey, predators would learn to avoid these individuals in subsequent encounters, and these individuals would be left untouched. Later Poulton (1890) broadened the de®nition from bright coloration to include any conspicuous warning signal, such as odours and sounds, and introduced the term `aposematism' (a0 po0 ; away, rgla; sign) to 606 distinguish it from crypsis (jqmpso0 q; hidden), the other end of the conspicuousness continuum. Conspicuousness is a common feature of aposematic prey, although there might be some exceptions (see Joron and Mallet, 1998). Typical warning colours such as red, yellow, orange and white, are usually associated with black patterning. These colours give the maximum visibility against brown and grey backgrounds (Cott, 1940), although the conspicuousness of aposematic prey has rarely been measured against their natural backgrounds. Since aposematic prey are advertising their unpro®tability to potential predators, the receiver is very important in this signalling system. It should be considered that the way in which receivers perceive colours and contrast is likely to be very dierent from that of human vision when we test the conspicuousness of aposematic prey (Osorio et al., 2001). For example, larvae of the grey shoulder knot (Lithophane ornitopus) are cryptic against their natural habitat of oak leaves in the visible spectrum, but not in the ultraviolet (UV) region (Church et al., 1998). Conspicuousness in the UV region does not necessarily reveal aposematism (see Lyytinen et al., 2001), but this requires further study. When studying the evolution of conspicuousness, the ®rst question is whether or not conspicuousness is an adaptive feature of warning coloration (see detailed discussion in Guilford, 1988, 1990). It is dicult to know whether conspicuousness is just the by-product of the coloration or whether it is an important component of warning signals. There are basically three experimentally studied hypotheses drawn from predator psychology proposed to explain why it is more bene®cial for unpalatable prey to use conspicuous colours rather than cryptic ones as a sign of unpro®tability (for more hypotheses, see Endler, 1991): 1. Predators learn to avoid unpalatable prey more rapidly if they are conspicuous rather than cryptic (Gittleman and Harvey, 1980; Gittleman et al., 1980; SilleÂn-Tullberg, 1985a; Roper and Wistow, 1986; Roper and Redston, 1987; Alatalo and Mappes, 1996; LindstroÈm et al., 1999a, 2001a). 2. Predators remember this association for longer (Roper, 1994). 3. Predators make fewer recognition errors if the pattern is conspicuous (Guilford, 1986). These ideas propose that aposematism as a defensive strategy has exploited the fact that predators from dierent taxa are capable of learning and remembering what is not bene®cial to eat and, therefore, it is an adaptive strategy against educated predators (see, e.g. Guilford, 1988, 1990). Since predators learn to avoid an unpro®table prey more quickly if it is conspicuous rather than cryptic, the evolution of aposematism represents a paradox: how can natural selection provide an advantage to a mutation that shifts a lifestyle initially from being cryptic to being highly conspicuous, since, surely, this 607 draws the attention of predators and attracts predation? How can this type of defence have evolved if predators sampled and killed the pioneers? Experimental approaches Surprisingly, although aposematism is one of the oldest evolutionary puzzles, the evolution of conspicuous signalling has attracted very few experimental studies. Perhaps this is due to the fact that conspicuousness is not considered to be as important as colours and odours, and that conspicuousness is only a byproduct of these characters. It is also very dicult to quantify conspicuousness, and therefore we are also lacking this data from the ®eld. Wickler (1968) thought also that it is impossible to study the initial origin of aposematism, and that we would learn more about warning colours by studying mimicry systems. In addition, the initial evolution happened so long ago that it might be fruitless to study this process (Mallet and Singer, 1987), which is such a rare event (Turner and Mallet, 1996). Nevertheless, although we might not be able to study the natural appearance of conspicuous aposematic prey, we can study the relative selection pressures of dierent aspects of conspicuous aposematic prey and draw hypotheses from these experiments to explain the evolution of conspicuousness in aposematic signalling species. However, we need to be careful in our experimental approach. First predator and prey species have co-evolved and might have multiple mechanisms to survive an encounter. These mechanisms, e.g. foraging preferences can confound our measures, for instance, of conspicuousness and might not reveal its true costs (see LindstroÈm et al., 2001a). Second, in order to separate parsimonious mechanisms, we need to compare the relative survival values of the aposematic prey to an evolutionarily reasonable alternative. For example, in the case of conspicuousness the comparison should be made between conspicuous unpro®table and cryptic unpro®table prey, to see whether aposematism is a more advantageous strategy, and if so, why (Guilford, 1988, 1990). Finally, we also need to be careful when we generalise our experimental results to the mechanisms explaining the evolution of aposematism. This is important if we draw our conclusion from experiments that are confounded by, for instance, predator biases. If predators have unconditioned aversion towards certain colours (see `Innate biases of the predators'), our experiments with these colours are likely to reveal more about the present world than give insights to the evolutionary mechanisms. I will ®rst discuss the problems of coevolution (`Innate biases of the predators' and `Adaptations of live insect prey to encounters with predators') and its impact on our experiments. Secondly, I will draw attention to use of alternative methods (`Arti®cial prey') namely the advantage of using arti®cial 608 prey to overcome these co-evolutionary adaptations. Finally I will discuss how the studies which real insects prey can provide interesting insight to the evolution of aposematism. Innate biases of the predators Cott (1940) and Edmunds (1974) argued that predators have no innate aversion against warningly coloured prey, and that each predator has to learn the association between a conspicuous signal and its unpro®tability. However, it has now been shown that predators do have unconditioned biases against typical warning colours (Smith, 1975, 1977; Schuler and Hesse, 1985; see LindstroÈm et al., 1999b and references therein). Predators might also have hidden colour biases, which can only be revealed when a second component of the warning display, for example pyrazine odour, is present (Rowe and Guilford, 1996, 1999, 2001). It seems that conspicuousness itself also becomes aversive in the presence of pyrazine, indicating that perhaps in multimodal signalling systems conspicuousness would not necessarily be initially costly (LindstroÈm et al., 2001b). Predator colour preferences are heritable, and birds can be arti®cially selected on the basis of their preferences (Kovach, 1978, 1983; Marples and Brake®eld, 1995). Biases can also be culturally transmitted: parents do not feed their nestlings with aposematic prey, which might result in a preference for cryptic prey and a bias against aposematic prey in the ¯edglings. Food preferences can also be learned from parents, who typically show their young what is bene®cial to eat rather than what to avoid (Turner, 1964; Avery, 1996). For aposematic prey such biases are bene®cial, since naõÈ ve predators will not be so naõÈ ve when they ®rst encounter aposematic prey. Although naõÈ ve birds may experiment and attack warningly coloured prey, colour biases may cause hesitation and allow prey to be sampled, or may subsequently speed up avoidance learning. These visual biases in predators are problematic in experiments that explore both the bene®ts of conspicuousness itself and the initial evolution of aposematism. For instance, conspicuousness might not be necessary for enhancing avoidance learning but the colour red is (SilleÂn-Tullberg, 1985b; Roper, 1990). It is, therefore, very dicult to separate the eects of pure conspicuousness from those of colour unless we control for coloration in our experiments. If we oer naõÈ ve predators a choice between edible mealworm larvae and a yellow and black monarch butter¯y larvae presented on a brown background, the results are likely to be confounded by foraging biases. If the predators readily eat all the brown larvae but only a few attack the conspicuous monarch butter¯y larvae, what can we conclude about conspicuousness? The results would tend to indicate that there is no cost of being conspicuous, but can we really conclude that there isn't a high visibility risk for the conspicuous aposematic 609 prey? And if the monarch butter¯ies survive predator attacks, can we then conclude that conspicuousness promotes prey survival, since conspicuousness makes birds initially more hesitant to attack? This result, where all brown prey are eaten and only on a few occasions do predators eat aposematic prey leads to the more interesting question of why predators only attack very few aposematic prey. The answer may lie in the foraging preferences of the predators: if predators prefer brown food over yellow food, then by using these colours in our experiments we might overestimate the survival values of unpalatable yellow prey, since the predator already has a tendency to avoid this colour. This unconditioned aversion towards the colour could be additionally reinforced by the unpro®tability of the prey, and thus the predator might only eat a single monarch larva. One way to overcome this problem experimentally is to have the colours reversed, for example, brown indicating unpalatability and yellow indicating palatability (see, e.g. Brower, 1960). However, given the diculties of accurately reversing the coloration of live prey, experiments often resort to the use of arti®cial prey. The advantage of using arti®cial prey is that we can more easily control for the predatory biases towards colours as well as manipulate conspicuousness. For example, we can use symbols instead of colours and test the selective advantages of dierent levels of conspicuousness. I will discuss this at length in `Arti®cial prey'. Adaptations of live insect prey to encounters with predators An additional problem with studying the initial origin of conspicuous signalling using real aposematic insect prey is that these prey species might have evolved other mechanisms to survive encounters with predators. These adaptations might be characters which we, as unnatural predators of these aposematic individuals, may consider meaningless. For instance, some insects use multicomponent signalling, where odours, sounds and behaviour are associated with conspicuous coloration (Cott, 1940; Edmunds, 1974); each of these components can interfere with conspicuousness. A moving aposematic insect might enhance avoidance learning by a predator, but then conspicuousness is only one part of aposematic strategy. It has been shown that the survival of an aposematic insect is best when all these components are presented simultaneously, but some characters like coloration in the defence is more important than others (Marples et al., 1994). Thus, when we use real insect prey to test the costs and bene®ts of conspicuousness, we might introduce adaptations that have an impact on the result unless we control for them. Therefore, interpreting our measures of survival in such experiments might be problematic. A good example is a tough cuticle that some insects have, which allows them to survive even when attacked (JaÈrvi et al., 1981; SilleÂn-Tullberg et al., 1982; Marples et al., 1994). Given that a toughened cuticle can enhance survivorship 610 of prey, we need to know: (1) How did the tough cuticle evolve? (Endler, 1991) (2) Can predators taste the unpalatability or noxiousness without eating the prey? (Kassarov 1999) (3) Are there any interactions between coloration and toughness, i.e. would aposematically coloured individuals with a tough cuticle be more easily avoided than if they were brown? Ideally we should answer these questions prior to suggesting, for instance, a mechanism (see Evans, 1987) for the evolution of aposematic signalling. However, in the absence of these data, we should ensure that when we measure the role of conspicuousness we are not confounding the results with either predator biases or adaptations of living insects. Arti®cial prey Since predators have unlearned preferences and live insect prey might have multiple defence strategies, applying the results of experiments using live prey or traditional warning colours (e.g. red, yellow, orange, unless the colours are controlled for) directly to the puzzle of the initial or even secondary evolution of conspicuousness is problematic to say the least. Therefore, to study the predation pressures that make conspicuousness a more favourable strategy than crypsis, it is necessary to conduct experiments where biases for certain colours and prey adaptations are taken into consideration. The bene®ts of using an arti®cial system is that it is possible to control for the adaptations of live insect prey, and it is also easier to control for the predator biases against certain colours. An additional bene®t is the ability to create dierent levels of conspicuousness. One powerful experimental method to study the bene®ts of conspicuousness, as well as the initial evolution of aposematism, is the use of a `novel world' ± a method created by Alatalo and Mappes (1996). This approach attempted to overcome the problems of predator biases by using a system where both the background `environment' and prey `species' consisted of arti®cial symbols. The landscape consisted of crosses, and prey animals were either cryptic, in that they had the background symbol printed on them, or they were made to contrast with the background by having a dierent symbol. By introducing dierent symbols on the prey, it was possible to manipulate the level of conspicuousness of the aposematic prey. The major advantage of this method is that it was now possible to use wild birds and regard them as evolutionarily naõÈ ve predators. In this way we can begin to assess the possible selective forces naõÈ ve predators might have generated to newly evolving conspicuous unpalatable prey. An obvious disadvantage is that we cannot directly relate selection pressures uncovered in these experiments to coloration, since it may be that some colours more easily account for the evolution of aposematism without the costs of conspicuousness (for example, the colour red, SilleÂn-Tullberg, 1985b). 611 Real insect prey Experiments with real insect prey can provide us with valuable insights into the costs and bene®ts of existing aposematic defence strategies. There are at least two fascinating systems where prey can have two morphs, one with a conspicuous signal and another with a more cryptic signal, indicating that there are dierences in predation or there is a cost of producing the conspicuous signal (see SilleÂn-Tullberg, 1985a, b; Sword, 1999). Lygaeus equestris bugs can be either grey or bright red: the red form has a selective advantage over the grey morph since predators are more hesitant to attack red bugs and predators also learn to avoid red individuals more rapidly than grey morphs (SilleÂn-Tullberg, 1985a, b). The grasshopper Schistocerca emarginata has a yellow and black morph that lives gregariously at high densities which also has an advantage over the low-density cryptic morph in relation to predation (Sword, 1999). These dierences allow us to look more closely into the bene®ts and costs of producing the conspicuous signal, which cannot be studied with arti®cial prey. These examples also highlight the potential importance of considering the life-histories of aposematic prey, since the animals' coloration re¯ects the end result of a combination of dierent selection pressures. For example, a colour polymorphism of ladybirds might be the result of a balance between predation and parasitism (Losey et al., 1997), or perhaps it is dependent upon their larval diet and is thus a result of plant±herbivore interactions (Grill and Moore, 1998; see also Tullberg et al., 2000a). Of course, predation pressures might change between life-history stages of prey. Presumably by taking the life-histories into account it would also give insights into evolutionary pathways: for example, whether conspicuous coloration is selected for adaptive defence or whether it was initially selected for other functions (see Mallet and Singer, 1987; Guilford, 1988, 1990; Endler 1991; Turner and Mallet, 1996; McMillan et al., 1997; Summers et al., 1997; Mallet and Joron, 2000). Whether we should then invoke the theory of aposematism to explain these adaptations has been questioned (Guilford, 1988, 1990). Experimental results As we now know, it is not straightforward to study the evolution of aposematism or the bene®ts of conspicuousness. In this paper, I review only 11 experiments, which test the survival of conspicuous unpalatable prey (i.e. `aposematic' prey) against a cryptic alternative prey. There are several experiments which compare the survival of aposematic prey against that of conspicuous palatable prey, thus in these experiments conspicuousness rarely possess a risk since alternative prey is easy to ®nd (e.g. JaÈrvi et al., 1981; Wiklund and JaÈrvi, 1982; Tullrot and Sundberg, 1991; Gagliardo and 612 Guilford, 1993; Gamberale and Tullberg, 1996a, b, 1998; Mappes and Alatalo, 1997; Tullrot, 1998), and I have excluded them from this review. Moreover, if we are interested in revealing the mechanism, which can account for the evolution of conspicuous signalling, adding contrast to the experiments is no longer sucient. Instead, we should also compare the selective value of a conspicuous prey, for example, by making comparisons between solitary and aggregated prey, or comparing the survival of novel versus familiar conspicuous prey. Surprisingly, all our experimental evidence on the bene®ts of conspicuousness comes from laboratory experiments (Gittleman and Harvey, 1980; Gittleman et al., 1980; SilleÂn-Tullberg, 1985a, b; Roper and Wistow, 1986; Roper and Redston, 1987; Roper, 1994; Alatalo and Mappes, 1996; LindstroÈm et al., 1999a, 2001a; Tullberg et al., 2000b). Most of these experiments have used arti®cial prey (except SilleÂn-Tullberg, 1985a, b). Nearly all these experiments conclude that conspicuousness is initially costly, but that this is later bene®cial through more eective avoidance learning. Conspicuousness also seems to be more risky if the prey is a colour, which is not a traditional warning colour (Sillen-Tullberg, 1985b). This could indicate that there is nothing special about conspicuousness per se, but that it is the combination of colour and contrast that enhances the learning process in predators. Instead, perhaps conspicuousness acts as an attractor that draws the attention of the predator. On the other hand, whether predators could use conspicuousness as a signal for unpalatability and generalise this to other conspicuous prey irrespective of the colour, has not been experimentally tested. The major dierence among experiments that investigate avoidance learning in predators is whether they also include palatable prey. Some experiments have tested the learning function of predators in the absence of palatable prey (Gittleman and Harvey, 1980; SilleÂn-Tullberg, 1985a; Roper and Wistow, 1986; Roper and Redston, 1987; Roper, 1994; Tullberg et al., 2000a), whereas others have included palatable prey (Gittleman et al., 1980; SilleÂn-Tullberg, 1985b; Alatalo and Mappes, 1996; LindstroÈm et al., 1999a, 2001a). We should bear in mind that the interpretations of the results dier. In the absence of palatable prey, the predators need to learn not to eat, whereas in the presence of palatable prey predators need to discriminate, and learn what to eat as well as what not to eat. This dierence in experimental design can lead to dierences in results and in the interpretations (Alatalo and Mappes, 1996, 2000; Tullberg et al., 2000a). It should also be noted that all experiments except one (Alatalo and Mappes, 1996) have used a frequency of 50% conspicuous and 50% cryptic prey. Since it is likely that conspicuous aposematic prey evolved from cryptic ancestors and were initially very rare, the possible anti-apostatic selection forces on rare conspicuous prey could in¯uence our results and diminish or increase the costs of conspicuousness. This can only be modelled until we have more experimental data on predator behaviour towards rare conspicuous prey. 613 It is very dicult to draw conclusions from our experiments investigating the evolution of conspicuous aposematic prey. Our experimental evidence is too scarce, since most experiments have tested the costs and bene®ts of conspicuousness and not the evolution of aposematism. There are two hypotheses, which have been experimentally tested: either conspicuousness evolved gradually (Yachi and Higashi, 1998, 1999; LindstroÈm et al., 1999a) or gregariousness allowed major mutations in conspicuousness (Alatalo and Mappes, 1996). Gradual evolution requires that a slight mutation in conspicuousness is selectively advantageous compared to a cryptic prey. This could be possible if predators would learn to avoid slightly conspicuous prey, which would not suer from visibility risks but still gain from avoidance learning. Later this learned avoidance would be transferred to an avoidance of even more conspicuous prey and conspicuousness could evolve through peak-shift mechanisms (Yachi and Higashi, 1998). However, predators only learn to avoid rather conspicuous prey, and furthermore peak-shift only seems to be advantageous when predators have learned to avoid rather conspicuous prey (LindstroÈm et al., 1999a). Nevertheless, gradual evolution might be possible through drift when there is no cost of being unpalatable, with predators eating slightly conspicuous prey in a similar way to palatable prey (LindstroÈm et al., 1999a). Alternatively, unpalatability in cryptic prey might have selected for gregariousness as Fisher (1930) suggested. Once the gregarious lifestyle had been attained, conspicuous signalling might have been selected for, since conspicuousness risk is diluted between individuals in a group (Alatalo and Mappes, 1996, 2000; but see Tullberg et al., 2000a). Additionally predators learn faster to avoid gregarious rather than solitary aposematic prey (Gagliardo and Guilford, 1993, Gamberale and Tullberg, 1996, 1998), and thus more individuals survive encounters with predators, which might have been essential when aposematic prey have been very rare. However, the bene®ts of gregariousness have been under explored, since we do not have a measure of how conspicuousness works in groups of prey. It has also been suggested that gregariousness might act as an signal itself (Cott, 1940; Gamberale and Tullberg, 1996; but see LindstroÈm et al., 1999b), but this has not been studied experimentally. It has also been suggested that gregariousness has evolved in prey that is already aposematic (SilleÂn-Tullberg, 1988, 1993). However, there might be very strong selection against aggregations: for instance, although Lygaeus equestris bugs have a lower predation risk when aggregated (Tullberg et al., 2000b), in nature they are rarely seen in groups. Perhaps the cost from competition over resources might have consequently prevented a gregarious lifestyle (Bertram, 1978). Therefore, our phylogenetic evidence of the evolution of aposematism in butter¯y larvae might re¯ect this strong selection against aggregations, although in other species gregariousness might have been a prerequisite for aposematic signalling. Additionally there could also be 614 dierent selection pressures among dierent life-stages of a particular species. It has been suggested that gregariousness might explain larval aggregations but individual selection would explain why adults are conspicuous (SilleÂn-Tullberg et al., 1982). Until we know more about aposematic prey under natural conditions, our suggestions on the mechanisms on the evolution are purely conjecture. There are other factors that could have in¯uenced the initial evolution of aposematism. Novelty eects, which are known to aect the food choices of generalist predators, might play a role in the evolution of aposematism (LindstroÈm et al., 2001a; see Marples and Kelly, 2001). However, it has been pointed out that novel signals might be selected against in an aposematic signalling system: it is likely that frequency-dependent disadvantages on a new rare morph relative to the common aposematic signal would cause predators to experiment more readily on them (Mallet and Singer, 1987; Mallet and Turner, 1996). However, novel appearance might not be unfavourable, since novel conspicuous prey might bene®t from peak-shift mechanisms, where predators bias their avoidance towards more conspicuous prey than those which they have already learned to avoid (LindstroÈm et al., 1999a; Gamberale and Tullberg, 2000). Curiously, novelty eects do not appear to lead to more rapid avoidance learning (see LindstroÈm et al., 2001a). Our evidence of predatory responses towards novel, rare aposematic prey is very scarce, and we need more experimental data before we can fully understand their eects on aposematic signal evolution. It still remains relevant as Guilford (1989) wrote that we need more controlled laboratory experiments to study both the selective values of conspicuous aposematic prey and the evolution of aposematism. Alternatives to the experimental approach are studying the evolution of conspicuousness using phylogenetic analyses (of prey: SilleÂn-Tullberg, 1988, 1993; Tullberg and Hunter, 1996; Read and Nee, 1997; but also of predators, Motychak et al., 1999) or by mathematical modelling (Harvey et al., 1982; SilleÂn-Tullberg and Bryant, 1983; Leimar et al., 1986; Mallet and Singer 1987; Endler 1988; SilleÂnTullberg and Leimar, 1988; Leimar and Tuomi 1998; Yachi and Higashi 1998, 1999; Speed, 2000). In the future, we need to focus on the particular question we are asking, and whether our experiments, phylogenetic studies or models are confounded with e.g. predator biases or multiple defensive mechanisms of aposematic prey. Acknowledgements I would like to thank everybody who participated the Aposematism: past, present and future-meeting in JyvaÈskylaÈ University, which was supported by the 615 Academy of Finland and University of JyvaÈskylaÈ, but especially Johanna Mappes, Rauno Alatalo, Candy Rowe, and two anonymous referees for constructive critique and improvement of the manuscript. References Alatalo, R.V. and Mappes, J. (1996) Tracking the evolution of warning signals. Nature 382, 708± 710. Alatalo, R.V. and Mappes, J. (2000) Initial evolution of warning coloration: comments on the novel world method. Anim. Behav. 59, F1±F2. Avery, M.L. (1996) Food avoidance by adult house ®nches, Carpodacus mexicanus, aects seed preferences of ospring. Anim. Behav. 51, 1279±1283. Brower, J.V.Z. (1960) Experimental studies of mimicry. IV. The reactions of starlings to the different proportions of models and mimics. Am. Nat. 94, 271±282. Church, S.C., Bennett, A.T.D., Cuthill, I.C., Hunt, N.S. and Partridge, J.C. (1998) Does lepidopteran larval crypsis extend into the ultraviolet? Naturwissenschaften 85, 1±5. Cott, H.B. (1940) Adaptive Coloration in Animals. Menthuen and Co., Ltd, London, England. Edmunds, M. (1974) Defence in Animals. A Survey of Anti-predator Defences. Longman, New York, USA. Endler, J.A. (1988) Frequency-dependent predation, crypsis and aposematic coloration. Phil. Trans. R. Soc. Lond. B 319, 505±523. Endler, J.A. (1991) Interactions between predators and prey. In J.R. Krebs and N.B. Davies (eds) Behavioural Ecology. Blackwell Scienti®c Publications, Oxford, pp. 169±196. Evans, D.L. (1987) Tough, harmless cryptics could evolve into tough, nasty aposematics: an individual selectionist model. Oikos 48, 114±115. Gagliardo, A. and Guilford, T. (1993) Why do aposematic prey live gregariously? Proc. R. Soc. Lond. B 251, 69±74. Gamberale, G. and Tullberg, B. (1996a) Evidence for a peak-shift of predator generalization among aposematic prey. Proc. R. Soc. Lond. B 263, 1329±1334. Gamberale, G. and Tullberg, B.S. (1996b) Evidence on more eective signal in aggregated aposematic prey. Anim. Behav. 52, 597±601. Gamberale, G. and Tullberg, B.S. (1998) Aposematism and gregariousness: the combined eect of group size and coloration on signal repellence. Proc. R. Soc. Lond. B 265, 889±894. Gamberale, G. and Tullberg, B.S. (2000) Experienced chicks show biased avoidance of stronger signals. An experiment with natural colour variation in live aposematic prey. Evol. Ecol. 13, 579±589. Gittleman, J.L. and Harvey, P.H. (1980) Why are distasteful prey not cryptic? Nature 286, 149±150. Gittleman, J.L., Harvey, P.H. and Greenwood, P.J. (1980) The evolution of conspicuous coloration: some experiments in bad taste. Anim. Behav. 28, 897±899. Grill, C.P. and Moore, A.J. (1998) Eects of a larval antipredator response and larval diet on adult phenotype in an aposematic ladybird beetle. Oecologia 114, 274±282. Guilford, T. (1986) How do `warning colours' work? Conspicuousness may reduce recognition errors in experienced predators. Anim. Behav. 34, 286±288. Guilford, T. (1988) The evolution of conspicuous coloration. Am. Nat. 131, s7±s21. Guilford, T. (1989) Studying warning signals in the laboratory. In R.J. Blanchard et al. (eds) Ethoexperimental Approaches to the Study of Behavior. Kluwer Academic Publisher, Dordrecht, pp. 87±103. Guilford, T. (1990) The evolution of aposematism. In D.L. Evans and J.O. Schmidt (eds) Insect Defence: adaptive Mechanisms and Strategies of Prey and Predators. State University of New York Press, New York, pp. 23±61. 616 Harvey, P.H., Bull, J.J., Pemberton, M. and Paxton, R.J. (1982) The evolution of aposematic coloration in distasteful prey: a family model. Am. Nat. 119, 710±719. Joron, M. and Mallet, J.L.B. (1998) Diversity in mimicry: paradox or paradigm? Trend. Ecol. Evol. 13, 461±466 JaÈrvi, T., SilleÂn-Tullberg, B. and Wiklund, C. (1981) The cost of being aposematic. An experimental study of predation on larvae of Papilio machon by the great tit Parus major. Oikos 36, 267±272. Kassarov, L. (1999) Are birds able to taste and reject butter¯ies based on `beak mark tasting'? A dierent point of view. Behaviour 136, 965±981. Kovach, J.K. (1978) Color preferences in quail chicks: generalization of the genetic selection. Behaviour 65, 263±267. Kovach, J.K. (1983) Perceptual imprinting: genetically variable response tendencies, selective learning, and the phenotypic expression of colour and pattern preferences in quail chicks (C. coturnix japonica). Behaviour 86, 72±88. Leimar, O. and Tuomi, J. (1998) Synergistic selection and graded traits. Evol. Ecol. 12, 59±71. Leimar, O., Enquist, M. and SilleÂn-Tullberg, B. (1986) Evolutionary stability of aposematic coloration and prey unpro®tability: a theoretical analysis. Am. Nat. 128, 469±490. LindstroÈm, L., Alatalo, R.V., Mappes, J., Riipi, M. and Vertainen, L. (1999a) Can aposematic signals evolve by gradual change? Nature 397, 249±251. LindstroÈm, L., Alatalo, R.V. and Mappes, J. (1999b) Reactions of hand-reared and wild caught predators towards warningly colored, gregarious, and conspicuous prey. Behav. Ecol. 10, 317± 322. LindstroÈm, L., Alatalo, R.V., Lyytinen, A. and Mappes, J. (2001a) Predator experience on cryptic prey aects the survival of conspicuous aposematic prey. Proc. R. Soc. Lond. B (in press). LindstroÈm, L., Rowe, C. and Guilford, T. (2001b) Pyrazine odour makes visually conspicuous prey aversive. Proc. R. Soc. Lond. B 268, 159±162. Losey, J.E., Ives, A.R., Harmon, J., Ballantyne, F. and Brown, C. (1997) A polymorphism maintained by opposite patterns of parasitism and predation. Nature 288, 269±272. Lyytinen, A., Alatalo, R.V., LindstroÈm, L. and Mappes, J. (2001) Can ultraviolet cues function as aposematic signals? Behav. Ecol. 12, 65±70. Mallet, J. and Joron, M. (2000) Evolution of diversity in warning color and mimicry: polymorphism, shifting balance and speciation. Ann. Rev. Ecol. System. 20, 201±233. Mallet, J. and Singer, M.C. (1987) Individual selection, kin selection, and the shifting balance in the evolution of warning colours: the evidence from butter¯ies. Biol. J. Linn. Soc. 32, 337±350. Mappes, J. and Alatalo, R.V. (1997) Eects of novelty and gregariousness in survival of aposematic prey. Behav. Ecol. 8, 174±177. Marples, N.M. and Brake®eld, P.M. (1995) Genetic variation for the rate of recruitment of novel insect prey into the diet of a bird. Biol. J. Linn. Soc. 55, 17±27. Marples, N.M. and Kelly, D.J. (2001) Neophobia and dietry conservatism: two distinct processes? Evol. Ecol., this issue. Marples, N.M., van Veelen, W. and Brake®eld, P.M. (1994) The relative importance of colour, taste and smell in the protection of an aposematic insect Coccinella septempuncata. Anim. Behav. 48, 967±974. McMillan, W.O., Jiggins, C.D. and Mallet, J. (1997) What initiates speciation in passion-vine butter¯ies? Proc. Nat. Am. Soc. 94, 8628±8633. Motychak, J.E., Brodie, E.D. Jr and Brodie, E.D. III (1999) Evolutionary response of predators to dangerous prey: preadaptation and the evolution of tetrodotoxin resistance in garter snakes. Evolution 53, 1528±1535. Osorio, D., MikloÂsi, A. and Gonda, Zs. (2001) Visual ecology and perception of coloration patterns by domestic chicks. Evol. Ecol., this issue. Poulton, E.B. (1890) The Colours of Animals. Their Meaning and Use. Especially Considered in the Case of Insects. Kegan Paul, Trench, TruÈbner and Co., Ltd, London. Read, A.F. and Nee, S. (1997) Inference from binary comparative data. J. Theoret. Biol. 173, 99± 108. 617 Roper, T.J. (1990) Responses of domestic chicks to arti®cially coloured insect prey: eects of previous experience and background colour. Anim. Behav. 39, 466±473. Roper, T.J. (1994) Eects of novelty on taste-avoidance learning in chicks. Behaviour 125, 265±281. Roper, T.J. and Redston, S. (1987) Conspicuousness of distasteful prey aects the strength and durability of one-trial avoidance learning. Anim. Behav. 35, 739±747. Roper, T.J. and Wistow, R. (1986) Aposematic coloration and avoidance learning in chicks. Quar. J. Exp. Psychol. 38B, 141±149. Rowe, C. and Guilford, T. (1996) Hidden colour aversion in domestic chicks triggered by pyrazine odours of insect warning displays. Nature 383, 520±522. Rowe, C. and Guilford, T. (1999) Novelty eects in a multimodal warning signal. Anim. Behav. 57, 341±346. Rowe, C. and Guilford, T. (2000) Aposematism: to be red or dead. Trend. Ecol. Evol. 15, 261±262. Rowe, C. and Guilford, T. (2001) The evolution of multimodal insect warning signals. Evol. Ecol., this issue. Schuler, W. and Hesse, E. (1985) On the function of warning coloration: a black and yellow pattern inhibits prey-attack by naive domestic chicks. Behav. Ecol. Sociobiol. 16, 249±255. SilleÂn-Tullberg, B. (1985a) Higher survival of an aposematic than of a cryptic form of distasteful bug. Oecologia 67, 411±415. SilleÂn-Tullberg, B. (1985b) The signi®cance of coloration per se, independent of background, for predator avoidance of aposematic prey. Anim. Behav. 33, 1382±1384. SilleÂn-Tullberg, B. (1988) Evolution of gregariousness in aposematic butter¯y larvae: a phylogenetic analysis. Evolution 42, 293±305. SilleÂn-Tullberg, B. (1993) The eect of biased inclusion of taxa on the correlation between discrete characters in phylogenetic trees. Evolution 47, 1182±1191. SilleÂn-Tullberg, B. and Bryant, E.H. (1983) The evolution of aposematic coloration in distasteful prey: an individual selection model. Evolution 37, 993±1000. SilleÂn-Tullberg, B. and Leimar, O. (1988) The evolution of gregariousness in distasteful insects as a defence against predators. Am. Nat. 132, 723±734. SilleÂn-Tullberg, B., Wiklund, C. and JaÈrvi, T. (1982) Aposematic coloration in adults and larvae of Lygaeus equestris and its bearing on muÈllerian mimicry: an experimental study on predation on living bugs by the great tit Parus major. Oikos 39, 131±136. Smith, S.M. (1975) Innate recognition of coral snake pattern by a possible avian predator. Science 187, 759±760. Smith, S.M. (1977) Coral-snake pattern recognition and stimulus generalization by naive great kiskadees (Aves: Tyrannidae). Nature 265, 535±536. Speed, M.P. (2000) Warning signals, receiver psychology and predator memory. Anim. Behav. 60, 269±278. Summers, K., Bermingham, E., Weigt, L., McCaerty, S. and Dahlstrom, L. (1997) Phenotypic and genetic divergence in three species of dart-poison frogs with contrasting parental behavior. J. Heredity 88, 8±13. Sword, G.A. (1999) Density-dependent warning coloration. Nature 397, 217. Tullberg, B.S. and Hunter, A.F. (1996) Evolution of larval gregariousness in relation to repellent defences and warning coloration in tree-feeding Macrolepidoptera: a phylogenetic analysis based on independent contrasts. Biol. J. Linn. Soc. 57, 253±276. Tullberg, B.S., Leimar, O. and Gamberale-Stille, G. (2000a) Did aggregation favour the initial evolution of warning coloration? A novel world revisited. Anim. Behav. 59, 281±287. Tullberg, B.S., Gamberale-Stille, G. and Solbreck, C. (2000b) Eects of food plant and group size on predator defence: dierences between two co-occurring aposematic Lygaeinae bugs. Ecol. Entomol. 25, 220±225. Tullrot, A. (1998) Evolution of Warning Coloration in the Nudibranch Polycera quadrilineata. D.Phil. Thesis, GoÈteborg, 26 pp. Tullrot, A. and Sundberg, P. (1991) The conspicuous nudibranch Polycera quadrilineata: aposematic coloration and individual selection. Anim. Behav. 41, 175±176. Turner, E.R.A. (1964) Social learning in birds. Behaviour 24, 1±46. 618 Turner, J.R.G. and Mallet, J.L.B. (1996) Did forest islands drive the diversity of warningly coloured butter¯ies? Biotic drift and the shifting balance. Phil. Trans. R. Soc. Lond. B 351, 835±845. Wallace, A.R. (1867) In Transactions of the Entomological Society of London, 3rd series, vol. III, London, 1864±1869. J. Proc. Entomol. Soc. Lond. Meeting, 4 March, 5, lxxx±lxxxi. Wickler, W. (1968) Mimicry in Plants and Animals. Weinsfeld and Nicolson Ltd, London. Wiklund, C. and JaÈrvi, T. (1982) Survival of distasteful insects after being attacked by naive birds: a repraisal of the theory of aposematic coloration evolving through individual selection. Evolution 36, 998±1002. Yachi, S. and Higashi, M. (1998) The evolution of warning signals. Nature 394, 882±884. Yachi, S. and Higashi, M. (1999) Modeling associative learning with generalization for a case of warning signals. Ecol. Res. 14, 243±248.